Year: 2025
Type: article
Abstract: This study presents an AI-powered network traffic detection framework capable of recognizing anomalies and addressing cyber threats in real-time. Traditional detection systems struggle to keep pace with evolving threats, necessitating more adaptive and intelligent approaches. To this end, the research integrates Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) models to enhance detection accuracy and operational efficiency. The framework is evaluated using benchmark datasets such as UNSW-NB15 and CICIDS2017, focusing on performance metrics including accuracy, precision, recall, and false positive rate. Experimental results show the proposed hybrid model achieves a detection accuracy of 92.08%, with precision and recall exceeding 92%, and a low average detection latency of 0.00142 seconds per sample. These findings confirm the model's effectiveness in detecting both known and novel threats, making it a scalable and reliable solution for modern cybersecurity challenges. The system offers real-time threat mitigation and valuable insights for network administrators, contributing to more proactive and robust security postures.
Institution: University of Lay Adventists of Kigali
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4411121887